Olfactory identification deficits occur in patients with Alzheimer's disease (AD), are associated with disease severity, predict conversion from mild cognitive impairment (MCI) to AD and are associated with healthy elderly subjects developing MCI. Odor (olfactory) identification deficits may reflect degeneration of cholinergic inputs to the olfactory bulb and other olfactory brain regions. Acetylcholinesterase inhibitors (ACheI) like donepezil show modest effects in improving cognition but can be associated with adverse effects and increased burden and costs because of the need for prolonged, often lifelong, treatment. Converging findings on odor identification test performance (UPSIT, scratch and sniff 40-item test) from four pilot studies, including two of our own, suggest that acute change in the UPSIT in response to an anticholinergic challenge (atropine nasal spray), incremental change over 8 weeks, and even the baseline UPSIT score by itself, may predict cognitive improvement with ACheI treatment in MCI and AD. If change in odor identification deficits can help to identify which patients should receive ACheI treatment, this simple inexpensive approach will advance the goal of improving personalized treatment, improve selection and monitoring of patients for ACheI treatment, reduce needless ACheI exposure with risk of side effects, and decrease health care costs. Project goals are to evaluate change in odor identification deficits as predictors of improvement in 100 patients with MCI (Study 1) and in 100 patients with mild to moderate AD (Study 2) during open treatment with the ACheI, donepezil. In both studies, we hypothesize that the decrease in UPSIT scores from pre- to post-atropine nasal spray challenge conducted at baseline will be associated with cognitive and global improvement from baseline to both 26 weeks and 52 weeks of donepezil treatment. The rationale is that those MCI patients who have AD brain pathology will have a cholinergic deficit and worsening of odor identification with the acute atropine challenge, and these patients will be more likely to improve with ACheI treatment. In both studies, we hypothesize that the increase in UPSIT scores from baseline (pre-atropine) to 8 weeks of donepezil treatment will be associated with cognitive and global improvement from 0 weeks to both 26 weeks and 52 weeks of donepezil treatment. This is the first systematic effort to evaluate the clinical utility of short-term changes in odor identificatin deficits to predict long-term cognitive and global improvement on ACheI treatment, and includes the first use of an anticholinergic challenge that has a sound neurobiological basis.